{
  "schemaVersion": "1.0.0",
  "benchmarkId": "humaneval@v1.0",
  "name": "HumanEval",
  "unit": "pass@1",
  "provenance": "ci-reproduced",
  "methodologyUrl": "/benchmarks/humaneval-v1/",
  "sourceUrl": "https://github.com/openai/human-eval",
  "notes": "Python function-completion benchmark (164 problems). Gaia reproduces via CI harness; each row carries a datasetHash + benchmarkInputHash. ci-reproduced rows count toward Trust Magnitude. pending rows are awaiting first CI reproduction run.",
  "rows": [
    {
      "skillId": "addy-osmani/code-simplification",
      "score": 0.75,
      "unit": "pass@1",
      "provenance": "ci-reproduced",
      "attestor": "https://github.com/gaia-research/gaia-skill-tree/actions/runs/28785817438@84bd47aa0fa1ae8bfb6de81b12cf1bc0674880b1",
      "datasetHash": "1d49078ba3e2b196b9344535bef34a43021f038fad9561d6ee7c53450609a6a2",
      "benchmarkInputHash": "020c89e15872648ee880ebccf6e6e67eb1ceae74f3f52aba4602fa61d417e2f7",
      "runAt": "2026-07-06T10:44:08Z",
      "harnessUrl": "https://github.com/gaia-research/gaia-skill-tree/blob/84bd47aa0fa1ae8bfb6de81b12cf1bc0674880b1/scripts/benchmarks/humaneval/run.py",
      "percentile": null,
      "modelRef": null,
      "notes": null
    },
    {
      "skillId": "addy-osmani/code-simplification",
      "score": 0.5,
      "unit": "pass@1",
      "provenance": "pending",
      "attestor": "pending-ci-reproduction",
      "datasetHash": "244753b2a3366bfbb271e76205fdd88e939c91705093c1a18eebd60fc8a0ebf8",
      "benchmarkInputHash": "3391b5f75da98f71962896b44acbfc37b37648d35474a10610e12b00c9e582a9",
      "runAt": "2026-07-05T00:00:00Z",
      "harnessUrl": "https://github.com/gaia-research/gaia-skill-tree/blob/286e46e72631bdb1e2332b3a9745255b3ddd0bda/scripts/benchmarks/humaneval/run.py",
      "percentile": null,
      "modelRef": null,
      "notes": "Dogfood seed; will be promoted to ci-reproduced by the first real workflow run of benchmark-humaneval-ci.yml."
    }
  ]
}